About

Ora is a scoring and auditing platform for what it calls agent experience: how easily AI agents can find, read, and use a website. As AI assistants increasingly decide which products to recommend or integrate with, many sites remain difficult for agents to navigate, with broken documentation links, missing pricing pages, and no machine-readable index. Ora measures this directly by scanning public URLs and grading the result.

Users enter a domain or MCP server URL, and Ora returns a 0-100 score with an A-F grade and a layered breakdown of what agents encounter. A companion feature called Journey points a live agent at a site with a chosen intent, such as "find pricing," and records every step, stall, and outcome. The product is built by era labs, and its scoring and API are accessible through a web dashboard, REST endpoints, MCP servers, and a command-line client published on npm as ax.

Key Features

Site and MCP scanning: Ora scans a domain, an MCP server URL, or an MCP App and returns a 0-100 score, an A-F grade, and a four-layer breakdown of agent readiness. Scores are computed from public URLs, so they reflect what a site actually exposes to agents.

Agent journeys: Journey sends a real agent to complete a curated task against a domain and records the full trajectory, including what it found, where it stalled, and whether it finished, with results streamed via SSE.

Leaderboards by category: Public rankings cover consumer, e-commerce, travel, developer tools, infrastructure, and agent-tool sites, allowing teams to compare their agent readiness against peers.

CLI auditing for coding agents: The `npx ax audit` command scores any URL from the terminal and prints ranked fixes; `--min-score` can gate CI pipelines, and `--tunnel-cmd` supports auditing localhost through a tunnel.

Open API and MCP interfaces: Reads are open without an API key, covering cached scores, fresh scans, discovery by intent, agent feedback, and journey runs through REST and MCP endpoints.

AgentReady open standard: Ora maintains AgentReady.org, an open standard for agent-ready products, developed with partners including Vercel, Stripe, Auth0, Mintlify, Tavily, Google, Elementor, and MIT.

Capability directory: A paid-capability index lists pay-per-call API endpoints with per-call USD prices and no API key required, searchable through the `search_capabilities` MCP tool or the directory.

Pricing & Fees

Pricing: Not publicly listed on the pages we reviewed; check the official site for current plans. Access to Ora's fix service is being rolled out slowly via a waitlist.
Public research: Ora publishes open research on web-wide agent readiness; it reports an average site score of 42 out of 100, with only 6% of sites scoring B or above.
Read access without a key: All read endpoints, including score lookups, leaderboards, and discovery, require no API key; only feedback writes are restricted and require agent verification.
Scan behavior: A fresh scan takes roughly 30 seconds, while deeper checks using live agent and LLM evaluation resolve asynchronously afterward.

Who It's For

Ora is aimed at engineering and platform teams who want their sites, documentation, and APIs to work well for AI agents. Developer-tool vendors and infrastructure providers can benchmark their agent readiness on public leaderboards, while e-commerce and consumer brands can diagnose why agents struggle to find pricing or complete tasks. The CLI and API also make it practical for DevOps teams to embed agent-readiness checks into CI, and for coding agents to audit and fix sites autonomously.

FAQ

What is Ora?Ora is a platform that scores how well AI agents can find, read, and use a website, assigning a 0-100 score and A-F grade based on public URLs, and it can run live agent tests against a domain to reveal where agents get stuck.

Is Ora free to use?Read access to scores, leaderboards, and discovery requires no API key and is openly available, while pricing for any paid plans or the fix service is not publicly listed on the pages reviewed; the official site has current details.

What platforms and integrations does Ora support?Ora offers a web dashboard, REST API, primary and docs MCP servers, WebMCP site tools, a CLI client on npm (`ax`), and skills for coding agents, so it works from browsers, terminals, and agent environments.

What are Ora's limitations?Ora only scans public URLs and cannot access private, internal, or login-gated sites, and some deep checks that use live agent and LLM evaluation resolve asynchronously after the initial 30-second scan.

How does Ora handle data and privacy?Every score is computed from public URLs only, so results reflect what a product actually exposes to agents rather than private internal information.

About

Key Features

Pricing & Fees

Who It's For

FAQ

Reviews (0)

Would you recommend Ora? Leave a comment

Related Tutorials

Related Tools